UWB Occupant Detection Using Multi-Decision Analysis to Cut Memory

The patent extracts only the essential information from the complete reflection images by using decision criteria that analyze specific characteristics (such as position, size, shape) of detected objects. Instead of storing and processing the entire reflection image database, the system stores and processes only the extracted analysis statements and decision criteria, significantly reducing memory requirements while maintaining detection reliability.

The patent segments the occupant detection task into multiple independent decision units, each responsible for analyzing specific aspects of the reflection images (e.g., person detection, animal detection, luggage detection). Each decision unit processes data streams independently and creates analysis statements separately, allowing for more efficient memory usage and processing.

The patent applies preliminary action by pre-defining decision criteria and reference characteristics for different types of occupants (persons, animals, luggage) before actual detection occurs. The evaluation device stores these pre-established criteria rather than storing complete reflection images, and uses them to quickly analyze data streams in real-time, reducing processing complexity while maintaining accuracy.

Instead of storing complete reflection images (which would require large memory and processing resources), the patent creates simplified copies or representations of the essential features - namely, analysis statements that capture the key characteristics of detected objects. These analysis statements serve as condensed representations that can be processed efficiently while maintaining the necessary detection accuracy.